Signed-off-by: Janne Grunau <janne-libav@jannau.net>

Signed-off-by: Diego Biurrun <diego@biurrun.de>

This reverts commit ac0fb593. It
causes valgrind errors which I'll want to investigate before
resubmitting this.

Signed-off-by: Mans Rullgard <mans@mansr.com>

We operated on 31-bits, but with e.g. lanczos scaling, values can
add up to beyond 0x80000000, thus leading to output of zeroes. Drop
one bit of precision fixes this.

When using e.g. lanczos scaling, values can drop below 0, so they
should never be unsigned.

We would use the second half of the U plane buffer, rather than the
V plane buffer, to output the V plane pixels.

For 9/10bit, it means we don't have to upscale to 16bit before
actual scaling or pixel format conversion, and thus a performance
gain.

This means that precision is retained when scaling between sample
formats with >8 bits per component (48bit RGB, 16bit grayscale,
9/10/16bit YUV).

Signed-off-by: Mans Rullgard <mans@mansr.com>

Signed-off-by: Mans Rullgard <mans@mansr.com>

This works through some non-obvious hacks in utils.c.

YUV planes were marked as uint16_t, but they contained signed data.
Fixes issue 1108 and 675.
Signed-off-by: Ronald S. Bultje <rsbultje@gmail.com>

Also remove code that overwrites the C versions of functions in
sws_init_swScale_altivec(), so that it uses the C functions of files
if no altivec-optimized version exists.

Fix handling of input if not in native endianness, and add support for
9/10-bit output. This allows us to force endianness of YUV420P 9/10bit
in the H264/10bit fate tests, which should fix them on big-endian
systems.

Signed-off-by: Anton Khirnov <anton@khirnov.net>

Signed-off-by: Anton Khirnov <anton@khirnov.net>

Signed-off-by: Anton Khirnov <anton@khirnov.net>

Instead of saving huge raw files, use the md5: output pseudo-protocol
to calculate the checksum of the file directly.  This is especially
useful when testing on remote targets as it avoids transferring 3.6GB
over the network.

code.

Originally committed as revision 26112 to svn://svn.ffmpeg.org/ffmpeg/trunk

Originally committed as revision 24806 to svn://svn.ffmpeg.org/ffmpeg/trunk

Increase readability and robustness, as the test result is not going
to differ if the order of the pixfmts codes changes.

Originally committed as revision 24665 to svn://svn.ffmpeg.org/ffmpeg/trunk

The corresponding lavfi-pixfmts BE tests are not yet added, as there
are some bugs in the scaler (scaling rgba, argb, bgra, abgr, yuva420p)
which result in differences with the LE reference, and I cannot
visually check the generated files on BE.

Originally committed as revision 24657 to svn://svn.ffmpeg.org/ffmpeg/trunk

This test verifies the pixdesc code by comparing the output with and
without a filter which should have no effect on the image.  Since the
available pixel formats depend on the byte order of the machine, a
simple reference checksum is not possible.

The test originally tried to solve this by generating a reference file
on the fly.  The problem with this is that the test framework expects
the reference file in the source tree, and writing to the source tree
is not allowed.

To avoid complicating the test framework, we instead provide two
reference files and select which to use based on the byte order.

Originally committed as revision 24330 to svn://svn.ffmpeg.org/ffmpeg/trunk

   Author: bcoudurier
   Date:   Sat Jul 3 03:11:04 2010 +0000

       Set graph swscale opts before parsing it, that way opts are available
       when auto-adding scalers.

It changed the swscale flags used by the auto-added scalers, and so
the output video.

Originally committed as revision 24065 to svn://svn.ffmpeg.org/ffmpeg/trunk

A patched version of ffmpeg supporting video filters is required for
getting this working; thus make lavfitest is supposed to work only in
the libavfilter repository for now.

Originally committed as revision 22586 to svn://svn.ffmpeg.org/ffmpeg/trunk